Mechanistic Insight into the Photoredox Catalysis of Anti-Markovnikov

Nov 12, 2014 - We describe our efforts to understand the key mechanistic aspects of the ..... Light on Unsaturated Hydrocarbons - “Gotta Heterofunct...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/JACS

Mechanistic Insight into the Photoredox Catalysis of AntiMarkovnikov Alkene Hydrofunctionalization Reactions Nathan A. Romero and David A. Nicewicz* Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290, United States S Supporting Information *

ABSTRACT: We describe our efforts to understand the key mechanistic aspects of the previously reported alkene hydrofunctionalization reactions using 9-mesityl-10-methylacridinium (Mes-Acr+) as a photoredox catalyst. Importantly, we are able to detect alkene cation radical intermediates, and confirm that phenylthiyl radical is capable of oxidizing the persistent acridinyl radical in a fast process that unites the catalytic activity of the photoredox and hydrogen atom transfer (HAT) manifolds. Additionally, we present evidence that diphenyl disulfide ((PhS)2) operates on a common catalytic cycle with thiophenol (PhSH) by way of photolytic cleaveage of the disulfide bond. Transition structure analysis of the HAT step using DFT reveals that the activation barrier for H atom donation from PhSH is significantly lower than 2-phenylmalononitrile (PMN) due to structural reorganization. In the early stages of the reaction, Mes-Acr+ is observed to engage in off-cycle adduct formation, presumably as buildup of PhS− becomes significant. The kinetic differences between PhSH and (PhS)2 as HAT catalysts indicate that the proton transfer step may have significant rate limiting influence.



INTRODUCTION Alkenes are one of the most versatile chemical feedstocks and are key components of innumerable synthetic transformations. A particularly active field of catalysis utilizes alkene reactants in hydrofunctionalization reactions such as olefin hydroalkoxylation and hydroamination reactions.1−3 A vast majority of these alkene hydrofunctionalization reactions proceed with Markovnikov selectivity. In the past decade and a half, there have been significant efforts by a number of research laboratories to develop catalytic protocols to access the opposite regioisomeric hydrofunctionalization adducts;4−6 however, a more general catalytic platform has yet to be identified. To address this, our laboratory has recently developed a number of methods for alkene hydrofunctionalization7−12 that have demonstrated the unique synthetic control accessible through systems which rely upon the well-defined redox cycles of a photoredox catalyst.13 These methods display complete antiMarkovnikov selectivity, employing a catalytic quantity of the organic dye 9-mesityl-10-methyl acridinium14−29 (Mes-Acr+)30 as a photooxidant along with a cocatalyst proposed to be a redoxactive hydrogen atom donor (Figure 1). One initial report from our group featured the use of MesAcrClO4 as a catalytic photooxidant along with 50−200 mol % 2phenylmalononitrile (PMN) as an H atom transfer (HAT) reagent in a hydroetherification reaction that proceeds with complete regioselectivity.7 This is particularly noteworthy in the context of oxidative alkene functionalizations, which often result in overoxidation and subsequent difunctionalization.31−34 Further optimization of this and related transformations identified thiophenol (PhSH) and, intriguingly, diphenyl disulfide ((PhS)2) as competent HAT catalysts, and these © XXXX American Chemical Society

Figure 1. Anti-Markovnikov hydrofunctionalization using Mes-Acr+ as a photoredox catalyst and PMN, PhSH, or (PhS)2 as viable HAT catalysts.

second-generation conditions have allowed for improved yields and drastically shortened reaction times. The increased efficiency rendered by arenethiol-based cocatalysts has enabled extension Received: June 20, 2014

A

dx.doi.org/10.1021/ja506228u | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

Scheme 1. Proposed Mechanism for Anti-Markovnikov Hydroetherification

catalytic systems is notoriously challenging. To understand the interdependent nature of dual catalyst cycles requires an in-depth inquiry beyond macroscopic study of overall rate and reaction order. Thus, we sought to conduct kinetic studies on the elementary steps in the proposed reaction mechanism toward elucidation of the rate limiting factors. We took a tandem approach in our study of the mechanism: steady state and transient absorption and emission spectroscopies were employed in determining rate constants for steps 1−2 and 5−6, while computational methods were utilized to offer complementary insight where spectroscopic study was impracticable (step 4).

of this anti-Markovnikov methodology to include a diverse array of nucleophiles, including carboxylic acids,8 amines,9,10 mineral acids such as HF, HCl, and MsOH,11 as well as propargylic and allylic alcohols and acids in a tandem addition-cyclization sequence.35,36 This demonstration of an efficient and broadly applicable complement to Markovnikov-selective protocols is a testament to the value of the alkene cation radical as an intermediate accessible via single electron transfer (SET). As these transformations are all believed to proceed by a similar mechanism, we were eager to establish a more intimate understanding of the reaction mechanism in order to further expand the synthetic utility of this reaction class. We viewed the intramolecular hydroetherification of alkenols as a model transformation for this study. Our current mechanistic hypothesis is depicted in Scheme 1, using alkenol hydroetherification as a representative example. Following single electron transfer from the alkene (1) to the electronically excited Mes-Acr+, the pendant alcohol undergoes intramolecular nucleophile addition to the alkenyl cation radical (2). Deprotonation of distonic cation radical 3 and subsequent hydrogen atom transfer (HAT) furnishes the cyclic ether (5). In the excited state, Mes-Acr+* is thought to undergo one electron reduction from the alkene; however, exciplex-mediated cyclization has been implicated in similar systems.37−43 The HAT catalyst is believed to operate in a concomitant redox cycle where HAT generates phenylthiyl radical (PhS·), which serves as a one electron oxidant for the acridine radical (Mes-Acr·). In this way, regeneration of ground state Mes-Acr+ and proton transfer to the resulting thiolate (7) completes a net redox-neutral cycle. The efficacy of the arenethiol-based HAT catalysts has been attributed in part to the oxidizing nature of PhS· (Ered 1/2 = +0.16 V vs SCE),44 which is expected to be an excellent redox partner for 11 oxidation of Mes-Acr· (Ered 1/2 = −0.55 V vs SCE). While many photoredox reactions feature additives that can greatly improve reaction efficacy through redox activity in parallel with the photosensitizer, few examples are truly catalytic with respect to the additive. In contrast, our system constitutes an interesting example where a redox active H atom donor seems to be catalytically relevant in both electron and proton transfer steps. However, mechanistic analysis of such multicomponent



RESULTS AND DISCUSSION Oxidative Activity of Excited State Mes-Acr+. To address the photocatalytic activity of Mes-Acr+, we focused on the use of transient spectroscopic methods. Although Mes-Acr+ has been a well-studied, yet contentious chromophore in recent years, photophysical studies have been mainly directed toward characterization of its excited state topology (Scheme 2). Verhoeven et al. report that the first singlet excited state of Mes-Acr+, localized on the acridinium system (hereafter referred to as the locally excited singlet state or LES) undergoes rapid intramolecular charge transfer from acridinium to the mesityl substituent to form the singlet CT state (CTS).26 LES and CTS are understood to be in thermal equilibrium, and fluorescence from both singlet states is measured on the nanosecond time scale. Moreover, both Fukuzumi and Verhoeven identify a longlived transient species that is observed to decay on the order of microseconds following laser excitation. Much of the debate has centered on the identity of this microsecond transient species, suggested by Fukuzumi to possess CT character and an excited state reduction potential (E*red) of +1.88 V vs SCE,14 while Verhoeven provides evidence that the species is the locally excited triplet state with E*red = +1.45 V vs SCE.26 In the absence of unambiguous evidence that the triplet state is comprised of two distinct states or that it is singly a CT or LE triplet, we will simply refer to this long-lived intermediate as the triplet (T), noting that T may denote CTT (charge transfer triplet) or LET (locally excited triplet), or both. B

dx.doi.org/10.1021/ja506228u | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

Table 1. Mes-Acr+ Fluorescence Quenching by Alkenes and HAT Catalysts

Scheme 2. Excited State Energy Diagram Adapted from Verhoeven26 and Fukuzumi15

In the course of our investigation, additional questions arose as to the photophysical nature of the excited state Mes-Acr+ in the midst of previous reports which draw varying conclusions from spectroscopic data. A crucial difference in our work was the use of nonpolar solvents such as 1,2-dichloroethane (DCE) rather than acetonitrile (MeCN), which was the medium employed in prior studies. Herein we share new evidence regarding the photophysical characteristics of Mes-Acr+ and its ET behavior in oxidation reactions with alkenes. Fluorescence Quenching: Rate of Primary Electron Transfer k1. Of the reports where Mes-Acr+ is used as a preparative photolytic oxidant, the long-lived transient (T) has been primarily implicated in inquiries of its excited state oxidative capacity.16,19−21,23,24,45 Although Fukuzumi presents evidence that T is responsible for oxidation of arenes with moderate oxidation potentials (e.g., anthracene; Eox = +1.19 vs SCE16), the oxidation potentials of many substrates employed in our methodology (e.g., 9−11, Table 1) approach or exceed the excited state reduction potential of T (E*red), which is estimated to lie between +1.45 and +1.88 V vs SCE based on the values reported by Verhoeven and Fukuzumi, respectively. Thus, while we acknowledged the possibility T could undergo reduction from more oxidizible alkenes (e.g., 8, 9, and 1b in Table 1 could be oxidized by CTT), it seemed unlikely that T-mediated oxidation could be general with respect to all alkenes used in our system, on the grounds that SET from alkenes 1 to T is endergonic in the cases where Ep/2 of the alkene exceeds +1.88 V. We considered the possibility that a viable pathway for oxidation is through SET to a singlet excited state of Mes-Acr+ (both LES and CTS are estimated to have excited state reduction potentials exceeding +2.0 V vs SCE).26 Since both singlet states are fluorescent, we elected to measure the rate of electron transfer by Stern−Volmer analysis of Mes-Acr+ fluorescence quenching.46 Employing Time-Correlated Single Photon Counting (TCSPC), we measure a fluorescence lifetime of 6.40 ± 0.03 ns for Mes-AcrBF4 in DCE.47 Stern−Volmer analysis was carried out on the observed quenching of fluorescence lifetime at increasing concentration of the quenchers given in Table 1. Anethole (8) quenches Mes-Acr+* most efficiently with a second

a KSV: Stern−Volmer Constant; error